llama.cpp

52d42686 - ci : add self-hosted-gpu-cuda and server-sanitize to hf-jobs (#28693)